GPU to 72 status... - Page 148

flashjh · 2012-12-05, 05:39

Quote:

Originally Posted by ixfd64

30 of mine were poached. I spent a whole week on them, so I'm not exactly too pleased about it.

I know George can manually apply credit, but still...

It wasn't on purpose - GPU72 gave the assignments out. Your TF assignments were only good for 30 days and after that they got reassigned. Since there is no way for GPU72 to remove the assignments from your worktodo files, you still finished them.

Logging into GPU72 to check your assignments page is the only way to see that they're expiring, expired or that someone else did the work.

Chris recently removed the 30 day expiration, for now, so this won't happen again.

ixfd64 · 2012-12-05, 05:40

The strange thing is that I requested them one week ago, which was when I started working on them.

flashjh · 2012-12-05, 05:43

Quote:

Originally Posted by ixfd64

The strange thing is that I requested them one week ago, which was when I started working on them.

I have it backwards then, it wasn't your expiration, it was his. You got them after they expired and then he still finished them (or reported completion) before you. He should have logged in and extended the reservations before they expired.

Edit: Again, Chris removed the 30 day expiration to prevent this from happening anytime soon.

chalsall · 2012-12-05, 13:11

Quote:

Originally Posted by ixfd64

30 of mine were poached. I spent a whole week on them, so I'm not exactly too pleased about it.

I know George can manually apply credit, but still...

I'm really sorry this happened. Duplication of effort is very frustrating to the participant who's work wasn't accepted, and is a waste of resources for the GIMPS effort; GPU72 was largely created to avoid this.

If it helps at all, you were credited with the work on GPU72, and as flashjh mentioned, at Garo's suggestion the expiry script has been disabled since we are now ahead of the LL (and DC) wave.

sonjohan · 2012-12-05, 17:10

Chris,

If at some point you neeed to reactivate it, maybe you could give a 'trusted user' switch for which the expiry doesn't apply.

It would avoid new guys to reserve stuff and enver send back information.

ixfd64 · 2012-12-05, 17:34

Quote:

Originally Posted by chalsall

I'm really sorry this happened. Duplication of effort is very frustrating to the participant who's work wasn't accepted, and is a waste of resources for the GIMPS effort; GPU72 was largely created to avoid this.

If it helps at all, you were credited with the work on GPU72, and as flashjh mentioned, at Garo's suggestion the expiry script has been disabled since we are now ahead of the LL (and DC) wave.

It's probably a moot point, but this is why people shouldn't bite off more than they could chew.

This remind me, would it be a good idea to configure PrimeNet to accept duplicate TF results as long as they're from different users? This would allow double checking for missed factors while preventing people from submitting results twice. I'd really like to hear George's thoughts on this.

LaurV · 2012-12-06, 02:15

Quote:

Originally Posted by ixfd64

This remind me, would it be a good idea to configure PrimeNet to accept duplicate TF results as long as they're from different users? This would allow double checking for missed factors while preventing people from submitting results twice. I'd really like to hear George's thoughts on this.

I won't exactly agree, except for P95 registered assignments, and P95 reports of those assignments (i.e. report is done by P95, with the w1 key). With mfaktX and CuLu the things are tricky, as there is no proof that we really did the assignment or just reported for credit. Not many people do P95 assignments for TF now since GPU's are so much faster, that would be a waste of resources, as CPU can be used more efficiently for DC and P-1. Allowing TF duplicates could be done either with "no credit" (just the database records) AND coming from different users, AND different programs maybe, or would involve implementing some w1 keys in mfaktx too. Even if the user is honest, a bug in mfaktc may duplicate the same "false negative".

flashjh · 2012-12-06, 05:23

@chalsall: On the View Assignments page under Production Heuristics, what is used for the 'GHz Days per Day' calculation? Can you make it so that the average is closer to the last two weeks or a month so the 'Days of Work' is more accurate?

petrw1 · 2012-12-22, 17:59

Recently GPU72 surpassed:
- 4,000,000 total GHZ days.
- 100,000 DC
- 100,000 LL
- 3,000,000 LL TF
- 500,000 DC TF
- 1,000,000 saved via LL TF
- 100,000 saved via DC TF
- 500,000 saved via P-1

petrw1 · 2012-12-23, 04:10

2012 - my year of DC is coming to an end.
My plan at the start of the year was to complete 1,000 DC.
Sadly, due to a variety of reasons I will come up short;
though I will complete about 825 by year end; 565 via GPUto72;
and close to 20,000 GhzDays.
I will move into the top 20 DC Lifetime (I did so recently)
and the top 10 DC for the year (just a couple days away).

2013 -Current plans, with Ivy and Sandy new to the team are to focus on LL
I will be putting at least 18 cores (all thru GPUto72) on LL including:
My new Ivy(4) and Sandy(2) along with i7(4), i5(4) and Q9550(4).
If I stick to this plan for all of 2013 I should acheive:
- over 25,000 GhzDays LL
- at least 250 completions
- top 100 LL Lifetime
- top 50 LL for the year

Hmmm looks like the horse and the dog are playing a game of "You get out of MY way; no you get out of MY way!!"

kracker · 2012-12-23, 17:23

Quote:

Originally Posted by petrw1

2012 - my year of DC is coming to an end.
My plan at the start of the year was to complete 1,000 DC.
Sadly, due to a variety of reasons I will come up short;
though I will complete about 825 by year end; 565 via GPUto72;
and close to 20,000 GhzDays.
I will move into the top 20 DC Lifetime (I did so recently)
and the top 10 DC for the year (just a couple days away).

2013 -Current plans, with Ivy and Sandy new to the team are to focus on LL
I will be putting at least 18 cores (all thru GPUto72) on LL including:
My new Ivy(4) and Sandy(2) along with i7(4), i5(4) and Q9550(4).
If I stick to this plan for all of 2013 I should acheive:
- over 25,000 GhzDays LL
- at least 250 completions
- top 100 LL Lifetime
- top 50 LL for the year

Hmmm looks like the horse and the dog are playing a game of "You get out of MY way; no you get out of MY way!!"

Hmm, I am just curious, how many DC's a day can you spit out with that kind of setup? Right now I myself do about 1-2 a day, will be pushed up soon though...

2012-12-22, 17:59	#1626
petrw1 1976 Toyota Corona years forever! "Wayne" Nov 2006 Saskatchewan, Canada 4692₁₀ Posts	A collection of recent milestones.... Recently GPU72 surpassed: - 4,000,000 total GHZ days. - 100,000 DC - 100,000 LL - 3,000,000 LL TF - 500,000 DC TF - 1,000,000 saved via LL TF - 100,000 saved via DC TF - 500,000 saved via P-1

2012-12-23, 04:10	#1627
petrw1 1976 Toyota Corona years forever! "Wayne" Nov 2006 Saskatchewan, Canada 2²·3·17·23 Posts	Personally... 2012 - my year of DC is coming to an end. My plan at the start of the year was to complete 1,000 DC. Sadly, due to a variety of reasons I will come up short; though I will complete about 825 by year end; 565 via GPUto72; and close to 20,000 GhzDays. I will move into the top 20 DC Lifetime (I did so recently) and the top 10 DC for the year (just a couple days away). 2013 -Current plans, with Ivy and Sandy new to the team are to focus on LL I will be putting at least 18 cores (all thru GPUto72) on LL including: My new Ivy(4) and Sandy(2) along with i7(4), i5(4) and Q9550(4). If I stick to this plan for all of 2013 I should acheive: - over 25,000 GhzDays LL - at least 250 completions - top 100 LL Lifetime - top 50 LL for the year Hmmm looks like the horse and the dog are playing a game of "You get out of MY way; no you get out of MY way!!"

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Status	Primeinator	Operation Billion Digits	5	2011-12-06 02:35
62 bit status	1997rj7	Lone Mersenne Hunters	27	2008-09-29 13:52
OBD Status	Uncwilly	Operation Billion Digits	22	2005-10-25 14:05
1-2M LLR status	paulunderwood	3*2^n-1 Search	2	2005-03-13 17:03
Status of 26.0M - 26.5M	1997rj7	Lone Mersenne Hunters	25	2004-06-18 16:46

2012-12-05, 05:40	#1619
ixfd64 Bemusing Prompter "Danny" Dec 2002 California 5×479 Posts	The strange thing is that I requested them one week ago, which was when I started working on them.

2012-12-05, 17:10	#1622
sonjohan May 2003 Belgium 2·139 Posts	Chris, If at some point you neeed to reactivate it, maybe you could give a 'trusted user' switch for which the expiry doesn't apply. It would avoid new guys to reserve stuff and enver send back information.

2012-12-06, 05:23	#1625
flashjh "Jerry" Nov 2011 Vancouver, WA 1,123 Posts	@chalsall: On the View Assignments page under Production Heuristics, what is used for the 'GHz Days per Day' calculation? Can you make it so that the average is closer to the last two weeks or a month so the 'Days of Work' is more accurate?